A Lightweight Conditional Diffusion Segmentation Network Based on Deformable Convolution for Surface Defect Detection
Surface defect detection is crucial to industrial manufacturing and research for surface defects has drawn much attention. However, defects in industrial environment are very diverse. Because defects scale and poses are constantly changing and current methods lack the ability to model the deformatio...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2025-01-01
|
| Series: | Journal of Electrical and Computer Engineering |
| Online Access: | http://dx.doi.org/10.1155/jece/2935790 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832542620527624192 |
|---|---|
| author | Jiusheng Chen Yibo Zhao Haibing Wang |
| author_facet | Jiusheng Chen Yibo Zhao Haibing Wang |
| author_sort | Jiusheng Chen |
| collection | DOAJ |
| description | Surface defect detection is crucial to industrial manufacturing and research for surface defects has drawn much attention. However, defects in industrial environment are very diverse. Because defects scale and poses are constantly changing and current methods lack the ability to model the deformation. To solve this problem, a lightweight conditional diffusion segmentation network based on deformable convolution is proposed. First, the conditional diffusion process is introduced for effective feature extraction; by gradually corrupting the defect images and recovering them from latent space, the model can obtain pixel-level segmentation results in an iterative process. Second, the efficient feature extraction block is proposed to address the problem of modeling varying defects, which is designed with a partial deformable convolutional layer that can fully extract geometric features of the diverse defects to further enhance the modeling power of the proposed network. Furthermore, the hyperparameters of the diffusion process are discussed to further improve the performance of the proposed method. The experimental results on DAGM2007, MT, AeBAD, and MVTec-AD indicate that the proposed model performs better than other baseline models. |
| format | Article |
| id | doaj-art-e90ac6519d7540c38425ff48c232c6c7 |
| institution | Kabale University |
| issn | 2090-0155 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Wiley |
| record_format | Article |
| series | Journal of Electrical and Computer Engineering |
| spelling | doaj-art-e90ac6519d7540c38425ff48c232c6c72025-02-04T00:00:03ZengWileyJournal of Electrical and Computer Engineering2090-01552025-01-01202510.1155/jece/2935790A Lightweight Conditional Diffusion Segmentation Network Based on Deformable Convolution for Surface Defect DetectionJiusheng Chen0Yibo Zhao1Haibing Wang2College of Electronic Information and AutomationCollege of Electronic Information and AutomationChina Southern Airlines Engineering and Technology Branch Beijing BaseSurface defect detection is crucial to industrial manufacturing and research for surface defects has drawn much attention. However, defects in industrial environment are very diverse. Because defects scale and poses are constantly changing and current methods lack the ability to model the deformation. To solve this problem, a lightweight conditional diffusion segmentation network based on deformable convolution is proposed. First, the conditional diffusion process is introduced for effective feature extraction; by gradually corrupting the defect images and recovering them from latent space, the model can obtain pixel-level segmentation results in an iterative process. Second, the efficient feature extraction block is proposed to address the problem of modeling varying defects, which is designed with a partial deformable convolutional layer that can fully extract geometric features of the diverse defects to further enhance the modeling power of the proposed network. Furthermore, the hyperparameters of the diffusion process are discussed to further improve the performance of the proposed method. The experimental results on DAGM2007, MT, AeBAD, and MVTec-AD indicate that the proposed model performs better than other baseline models.http://dx.doi.org/10.1155/jece/2935790 |
| spellingShingle | Jiusheng Chen Yibo Zhao Haibing Wang A Lightweight Conditional Diffusion Segmentation Network Based on Deformable Convolution for Surface Defect Detection Journal of Electrical and Computer Engineering |
| title | A Lightweight Conditional Diffusion Segmentation Network Based on Deformable Convolution for Surface Defect Detection |
| title_full | A Lightweight Conditional Diffusion Segmentation Network Based on Deformable Convolution for Surface Defect Detection |
| title_fullStr | A Lightweight Conditional Diffusion Segmentation Network Based on Deformable Convolution for Surface Defect Detection |
| title_full_unstemmed | A Lightweight Conditional Diffusion Segmentation Network Based on Deformable Convolution for Surface Defect Detection |
| title_short | A Lightweight Conditional Diffusion Segmentation Network Based on Deformable Convolution for Surface Defect Detection |
| title_sort | lightweight conditional diffusion segmentation network based on deformable convolution for surface defect detection |
| url | http://dx.doi.org/10.1155/jece/2935790 |
| work_keys_str_mv | AT jiushengchen alightweightconditionaldiffusionsegmentationnetworkbasedondeformableconvolutionforsurfacedefectdetection AT yibozhao alightweightconditionaldiffusionsegmentationnetworkbasedondeformableconvolutionforsurfacedefectdetection AT haibingwang alightweightconditionaldiffusionsegmentationnetworkbasedondeformableconvolutionforsurfacedefectdetection AT jiushengchen lightweightconditionaldiffusionsegmentationnetworkbasedondeformableconvolutionforsurfacedefectdetection AT yibozhao lightweightconditionaldiffusionsegmentationnetworkbasedondeformableconvolutionforsurfacedefectdetection AT haibingwang lightweightconditionaldiffusionsegmentationnetworkbasedondeformableconvolutionforsurfacedefectdetection |